120 



KANSAS ACADEMY OF SCIENCE. 



of all the various rays which enter into the composition of the colors in question, 

 give no definite knowledge of the nature of the light to which pigments owe their 

 characteristic hues. 



Soon after the completion of the measurements the results of which are indicated 

 in these four curves, it occurred to me that an important step toward the solution of 

 the problem of the color of the sky might be made by directing the spectropho- 

 tometer to the sky itself, and comparing its spectrum throughout with that of some 

 neutral white substance capable of reflecting all colors equally well. I attempted 

 such an analysis in the spring of 1885. It is not my purpose to enter here into the 

 details of that research. Some of the difficulties encountered in the selection of a 

 white substance suitable for the comparison will be described in my paper entitled 

 "Black and White." [See pp. 37-44 of these Transactions.] The pigment finally 

 adopted was the carbonate of magnesium. This analysis brought out the remark- 

 able fact that the light 

 which the sky reflects 

 to us, although differ- 

 ing from direct sun- 

 light to a measurable 

 extent, yet corre- 

 sponds with the latter 

 quite as closely as does 

 that reflected by any 

 white pigment which 

 I have studied. Its 

 deviation from true 

 whiteness, which va- 

 ries from day to day 

 and from hour to 

 hour, does not always 

 consist as the appear- 

 ance of the sky would 

 lead us to suppose, in any marked excess of blue rays. Indeed, the first sky which 

 I subjected to measurement was found to be deficient in the very rays upon which 

 the existence of an objective blue depends. The curves now shown upon the screen 

 are those pertaining to the spectrum of the sky on April 28, May 1. and May 4, 1885. 

 Light reflected from a perfectly white body would be represented by the horizontal 

 line. ( Fig. 5.) Deviations from that line denote excess or deficiency of the rays 

 belonging to the corresponding regions of the spectrum, as indicated by the Fraun- 

 hofer lines at the bottom of the diagram. 



On April 28 (see Curve 1, Fig. 5), the sky was of more than average blueness, so 

 far as the eye could judge, yet its spectrum was found to possess more red and less 

 blue and violet than belong to a true white, being in this respect almost identical 

 with the magnesium carbonate with which it was compared. On the first t>f May 

 (see Curve 2, Fig. 5), a sky, to the unaided eye very like the first, showed a decided 

 excess of blue; while on the 4th of May (see Curve 3, Fig. 5), the predominating color, 

 as shown by the spectrophotometer, was greenish-blue, the violet rays being com- 

 paratively feeble. 



In order to understand the extent to which these variations from true whiteness 

 were capable of influencing the hue of the sky as observed by the unaided eye, let 

 us compare these curves with those obtained by the same method and with the same 

 instrument from various substances, which had been selected because of their un- 

 usual whiteness. Specimens of magnesium carbonate, of plaster of Paris and of 



B C 



li b 

 Fig. 5. 



H 



